Qwen3.5-9B-Claude-4.6-Opus-abliterated
9BQwen
from Jackrong/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-v2
↓ 9K♥ 657 repos
Qwen3.5-9B-Claude-4.6-Opus-abliterated-heretic
Qwen3.5-9B-Claude-4.6-Opus-abliterated-heretic is a Heretic-ablated variant of Jackrong/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-v2, published on Hugging Face by sulpikar2. It is 9B parameters, single consumer GPU class and apache-2.0 licence. It is distributed across 5 repositories in Transformers, GGUF and GGUF (imatrix) formats, totalling 6.8K downloads.
Specification
- Model name
- Qwen3.5-9B-Claude-4.6-Opus-abliterated-heretic
- Source repository
- sulpikar2/Huihui-Qwen3.5-9B-Claude-4.6-Opus-abliterated-heretic
- Publisher
- sulpikar2
- Parameters
- 9B
- VRAM required
- ~6.2 GB at Q4_K_M, 4k context (estimated) — runs on 8 GB cards
- Hardware class
- 4B – 10B — single consumer GPU
- Licence
- apache-2.0
- Task
- Image text to text
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- Transformers, GGUF, GGUF (imatrix)
- First indexed
- 23 Mar 2026
- Last updated
- 14 Oct 2026
Ablation technique
How Qwen3.5-9B-Claude-4.6-Opus-abliterated-heretic was modified, and what that implies.
Heretic
Automated directional ablation via the Heretic toolchain, which searches for the refusal direction and applies it with a KL-divergence budget so general capability is preserved.
Reported scope: Not stated by the publisher. Publishers frequently omit this, so verify behaviour empirically rather than assuming full coverage.
From the publisher’s model card
“This is an uncensored version of Jackrong/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-v2 created with abliteration (see remove-refusals-with-transformers to know more about it). This is a crude, proof-of-concept implementation to remove refusals from an LLM model without using TransformerLens.”
Running it
Commands are templates — confirm the exact repository and quant file before use.
Ollama (publisher-documented)
ollama run huihui_ai/qwen3.5-abliterated:9b-Claudellama.cpp / GGUF
hf download mradermacher/Huihui-Qwen3.5-9B-Claude-4.6-Opus-abliterated-heretic-i1-GGUF --local-dir ./qwen3.5-9b-claude-4.6-opus-abliterated-heretic
llama-cli -m ./qwen3.5-9b-claude-4.6-opus-abliterated-heretic/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "sulpikar2/Huihui-Qwen3.5-9B-Claude-4.6-Opus-abliterated-heretic"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve sulpikar2/Huihui-Qwen3.5-9B-Claude-4.6-Opus-abliterated-heretic --trust-remote-codeDownloads and variants (4)
Every published repository of Qwen3.5-9B-Claude-4.6-Opus-abliterated-heretic, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| mradermacher/Huihui-Qwen3.5-9B-Claude-4.6-Opus-abliterated-heretic-i1-GGUF | GGUF (imatrix) | 3.4K |
| mradermacher/Huihui-Qwen3.5-9B-Claude-4.6-Opus-abliterated-heretic-GGUF | GGUF | 3K |
| fairy322/Huihui-Qwen3.5-9B-Claude-4.6-Opus-abliterated-heretic-GGUF | GGUF | 427 |
| sulpikar2/Huihui-Qwen3.5-9B-Claude-4.6-Opus-abliterated-hereticsource | Transformers | 26 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- Qwen3.5-9B-Claude-4.6-Opus-abliterated-heretic will attempt offensive-security prompts that a hosted commercial model declines, which is what makes it usable for red-team corpus generation and for measuring what an unaligned model of this class produces.
- Capability is inherited from Jackrong/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-v2, not added by ablation — at 9B parameters, expect a small model’s reasoning and a high error rate on exploit detail.
- The publisher does not state which layers were ablated, so assume nothing about refusal consistency — probe it directly.
- Declared licence is apache-2.0, but the terms that bind you are Jackrong/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-v2’s — a re-release cannot grant more than its parent.
Handling: run it isolated, put moderation in front of it if anyone outside your team can reach it, and verify the weights before loading — these are third-party artifacts. Full handling and licence guidance.
Metadata for Qwen3.5-9B-Claude-4.6-Opus-abliterated-heretic is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with sulpikar2, and does not independently verify publisher claims. See responsible use.